Ladders

that's what I'm saying it only works for 1v1, we need something that is applicable to 20 v 20

and yes that's basically why I said "their own"

no the most important thing we need rating for is 1v1. And if you have fixed teams you can apply elo to 20v20 as well

what???? dude 1v1 can't sum up pro PA, otherwise what are we fighting for and why arn't we passionate about starcraft 3 instead?

ELO is for chess this is the same as P2P F2P, it's not applicable to PA.

And I refuse to have different rating systems for different ladders

https://forums.uberent.com/threads/elo-vs-true-skill.49426

This opinion is going to draw a lot of ire, and will certainly be in the minority, but in the interest of providing a dissenting view, I think it needs to be said...

None.

A few problems I've seen in the past due to ladders:
1) In SupCom, the ladders caused a good deal of inter-community strife regarding game balance between "1v1 ladder" and "casual let's make some awesome explosions" players.

Balance requirements for 1v1 will likely be entirely different than for 20v20. Which group gets priority?
I've always seen the TA/SupCom/PA bloodline as the juxtaposition of Starcraft. (Focus on strategy and scale, by giving you the most powerful interface possible, that lets you implement grand ideas easily, instead of making the rapid use of the interface itself the centerpiece of the game). I'm not sure 1v1 really embodies this ideal, but it's the most logical "entry point" for a ladder system. These factors are at odds with one another.
​

2) Segregating an already small player-base is a great way to sabotage the lifespan of a game.

If you're going to have 1v1 and 2v2 and 3v3 and "custom game" and "ranked explosion fest" and countless other modes, the pool of players searching for any given permutation shrinks. Wait times for games increase. Maybe my list of game-types is hyperbole, but the concept stands. It can be very frustrating to begin a game for a player who's favorite game-type may not be the most popular. ​

3) Developer work-load.

If one of the "lower tier?" developers can implement ladders, then alright. But if one of the guys doing path-finding or performance tweaks needs to dedicate his valuable time to creating a ladder, then forget it. As has been mentioned numerous times, Uber is a small developer. The game's budget is small given the scope of the project. Is a ladder a "core" feature? Is the opportunity cost of it's development worth it? Beware feature-creep!
​

That said, I do believe some kind of hidden player skill metric would not be a bad thing to include, to facilitate random matchmaking.

how do you want this game to be an e-sport if there's no ladder?

I agree with the fact they can make them after the fact though

I think ladders could have overlaping game sizes. For example we could have 4 ladders for a different range of game sizes:

• 1v1
  
• 2v2
• 3v3 to 5v5
• 6v6 and larger

1) Such a divide between player preference will exist whatever you do. We already have a ladder set up by the community. Even without that we would have a community of competitive 1v1 players. I think the issues you raise in this point are nonexistent. It is possible to make a game that works from 1v1 to 10v10 (I have done so) and when done right the same balance applies to both sizes of games. It is reasonable to balance for 1v1 and then construct the game in such a way that the balance carries over to larger games. Also 1v1 games do not have to be about rapid use of the UI, in PA I hope that they are not.

2) Simply let players search for games of many different types simultaneously and put them in the first one available. Dota 2 does this.

3) I partially agree here. It is not a core feature. I think someone in the community would make a ladder system. I don't particularly mind if Uber works on it but if it is easy for them to do it would be a nice thing to have.

1v1, 2v2, 3v3 at least.

I love all the procedural stuff PA has going for it. i personally think uber should throw the whole "fair" map thing out the window as an outdated way of doing a ladder. what makes this game awesome and unique is the random planets and setups. you have more than one place to pick for starting on a given planet and if they expand this idea to picking a planet and then picking a spot i think you can mitigate the difference in starting spots.

Defining player skill should partially be based on picking your starting location and dealing with what you have vs what your opponent has.....what i mean is pick the best location based on the information you have and then scout your opponent and plan accordingly based on what he has. if he has a lot of metal right next to him either expand quickly or attck quickly.

giving both players the same thing and saying go doesnt to me encompass all the different skills it takes in PA to be good. in fact it takes one of the main skills out of the mix....adaptability.

GPGNET used ELO for 1vs1, 2vs2 and 3vs3.
I realy liked the system because it's an easy system were you always know how much points you can win. Also you know that you can't lose more then 32 points. FAF is using trueskill now. This system is also working great, but i can't say before a game starts how much points i can lose/win. It's too complicate for that. That doesn't mean it's bad.
So i have no problem if we will use trueskill for PA, but i personly prefer ELO.
(and yes i'm playing chess )

yeah the simplicity of elo is it's great advantage.

@tatsujb : 1v1 is THE kind of thing that matters. I so don't care for teamgames
@roadtoad42/googlefrog:
googlefrogs answers on 1 and 2 are pretty good. On 3 however I disagree. A build in ladder automatch system is very _VERY_ helpful to quickly find good games. Currently it's hard to get good games going because there are only few players who bother to join the irc and find matching opponents. Automating this process is a tremendous improvement that also, once online replays are available, allows anybody to quickly check out a few replays of good players.
A 1v1 automatch ladder system is a very fundamental feature of any RTS.

elo is inaplicable even to 1v1 in PA, too many parameters in PA, it's just not chess.

with faf the switch was made to trueskill and look how much of a bigger playerbase it has compared to gpgnet

elo won't help you get to the top and neither will trueskill hinder you doing so.

Don't you think that's kinda unfair or unfounded? Do you have any stats from GPGnet to support that?

Mike

Well this is just plain incorrect.

The number of parameters that lead to a win or loss is 100% irrelevant here. As long as a match can end up either with a win, draw or loss, ELO is applicable.

Not to mention that theoretically, ELO can be used for even random team matchmaking, with some tweaks. It has been done in other games(Zero-K uses elo in teamgames to great effect).

Honestly though, I'm not even sure what would be the best for PA at this point. Maybe Uber could do a poll on this? I'd be fine with ELO, but I also wouldn't mind a Starcraft 2 style system with hidden MMR and different visible leagues and divisions.

Whatever ranked is, please make it so you can't make a system, sick of joining custom game and some tit thinks it's funny to have 12 or so planets.

for me 1v1 symmetric maps sounds a bit retro, the game should probably also last only a few minutes, right?

how about x v x v x with x relatively small, given the community size it would probably be a viable alternative.
if the planet generation and metal distribution is improved, I don't see a reason for symmetric maps.

Aside from symmetric or not, can you guys deal with different solar system constellations or does it have to be a standard, fixed set of planets?

there are more variables than just the decision on the x in the case of x v x.
the solar system level alone adds a new dimension to the variations.

It doesnt have to be a fixed set of systems, it just needs to be balanced, which is why we need symmetry. Doing something new, or non-retro, is not always a good thing.

also I find it should be clarifed that all ladder matches should be played on randomly spawned systems with everything set to random, planet number, size, what orbits around what, and all settings relating to each planet.
this would allow to fully utilise the procedural generation engine and remove map memorizing, with in this game is worlds appart from FA where map memorizing only brought a punny advantage compared to PA.

no

why?

Maybe I should clarify: Only within limits. Depending on how well the generation turns out in the future I am certain that some limitations will be required. beginning with a simple "not too big, not too small" ending -if the generation does not improve a lot- with symmetrical planets.

Also do I correctly understand you that you're saying that map memorization gives a larger in advantage in PA when compared to FA?

yes
we have a backroud of such maps like "setons" where map memorization is 70%gameplay there are map strategies.
a whole contex to learn,

well i found that in PA with random maps, this context can get even bigger and make an even bigger difference between those who know the map and those who don't.
on my maps I am undefeatable.